The ACC/TVT score is a specific predictive model of in-hospital mortality for patients undergoing transcatheter aortic valve implantation (TAVI). The aim of this study was to test its predictive accuracy in comparison with standard surgical risk models (Logistic Euroscore, Euroscore II, and STS-PROM) in the population of TAVI patients included in the multicenter RISPEVA (Registro Italiano GISE sull'impianto di Valvola Aortica Percutanea) registry. The study cohort included 3293 patients who underwent TAVI between 2008 and 2019. The 4 risk scores were calculated for all patients. For all scores, the capability to predict 30-day mortality was assessed by means of several analyses testing calibration and discrimination. The ACC/TVT score showed moderate discrimination, with a C-statistics for 30-day mortality of 0.63, not significantly different from the standard surgical risk models. The ACC/TVT score demonstrated, in contrast, better calibration compared with the other scores, as proved by a greater correspondence between estimated probabilities and the actual observations. However, when the ACC/TVT score was tested in the subgroup of patients treated in a more contemporary period (from 2016 on), it revealed a slight tendency to lose discrimination and to overestimate mortality risk. In conclusion, in comparison with the standard surgical risk models, the ACC/TVT score demonstrated better prediction accuracy for estimation of 30-day mortality in terms of calibration. Nevertheless, its predictive reliability remained suboptimal and tended to worsen in patients treated more recently.
Copyright © 2020. Published by Elsevier Inc.